A 100 mpg car is highly constrained by the laws of physics to be crappy, or at least extremely unappealing.
A much better goal, in terms of usefulness, challenge, and well-defined-ness, would be to develop a self-driving car. This probably could have been done 20 years ago if we had started a Manhattan-style project. But I guess we'll just have to wait for Google.
These incremental advances have been appearing for at least 20 years, and eventually we will have self-driving cars even without an organized effort (either by the government or a company like Google with massive resources). But that's a very slow option.
A much better goal, in terms of usefulness, challenge, and well-defined-ness, would be to develop a self-driving car. This probably could have been done 20 years ago if we had started a Manhattan-style project. But I guess we'll just have to wait for Google.